Stat 313

New design lofar2_unb2b_adc with revisions one_node and full for lab testing with the ADC board.

New IP version e2sg for UniBoard2 with Quartus 19.4

Revision unb2c_test_pinning of design unb2c_test compiles Ok. Final pin locations are still to be determined

Hi Jonathan,

1. central quartus dir I think the quartus/ dir that you now have per revision directory in:

applications/lofar2/designs/lofar2_unb2b_adc/revisions/lofar2_unb2b_adc_one_node/quartus/ applications/lofar2/designs/lofar2_unb2b_adc/revisions/lofar2_unb2b_adc_full/quartus/

could be place central in the 'parent' library:

applications/lofar2/designs/lofar2_unb2b_adc/quartus

because contain the same. If one revision needs other quartus settings than another, then still the settings could be in the central quartus dir, because the hdllib.cfg per revision can make the appropriate differentiations.

For example see how we did it in ARTS, APERTIF:

https://svn.astron.nl/viewvc/UniBoard_FP7/RadioHDL/trunk/applications/apertif/designs/apertif_unb1_correlator/ https://svn.astron.nl/viewvc/UniBoard_FP7/RadioHDL/trunk/applications/arts/designs/arts_unb1_sc4/

2. Revisions parent instance is unb2b_test_lib.unb2b_test instead of lofar2_unb2b_adc.vhd I would expect that the revision instances the entity from the parent design. I also find it strange that it works at all, because the unb2b_test_lib is not mentioned at the LIBRARY clause in:

https://git.astron.nl/-/ide/project/desp/hdl/blob/stat-313/-/applications/lofar2/designs/lofar2_unb2b_adc/revisions/lofar2_unb2b_adc_full/lofar2_unb2b_adc_full.vhd https://git.astron.nl/-/ide/project/desp/hdl/blob/stat-313/-/applications/lofar2/designs/lofar2_unb2b_adc/revisions/lofar2_unb2b_adc_one_node/lofar2_unb2b_adc_one_node.vhd

3. Entity and package with the same name:

https://git.astron.nl/-/ide/project/desp/hdl/blob/stat-313/-/applications/lofar2/designs/lofar2_unb2b_adc/src/vhdl/lofar2_unb2b_adc.vhd https://git.astron.nl/-/ide/project/desp/hdl/blob/stat-313/-/applications/lofar2/designs/lofar2_unb2b_adc/src/vhdl/lofar2_unb2b_adc_pkg.vhd

Propose to rename the package into PACKAGE lofar2_unb2b_adc_pkg IS.

4. central testbench and revision test benches Note: in the revision design directory it is typically not necessary to use subdirectories, because there are only a few files. Note: for example see:

a) applications/lofar2/designs/lofar2_unb2b_adc/hdllib.cfg The central test bench is commented out in the hdlib.cfg, I think we should have at least a minimal testbench that just runs for 1 us to show that the toplevel design entity can compile, load and simulate. It does not need other stimuli than just the clock.

b) similar each revision can have a test bench to show that the HDL can compile, load and simulate for 1 us.

For example see: https://svn.astron.nl/viewvc/UniBoard_FP7/RadioHDL/trunk/applications/apertif/designs/apertif_unb1_correlator/revisions/apertif_unb1_correlator_full/tb_apertif_unb1_correlator_full.vhd

PS: We can communicate via these comments. If you make a change on the branch and push it I think I will automatically see it in this merge request. It is not necessary to renew this merge request, it remains pending. When the review is done (ie. both the coder and the reviewer are ok with it), then the reviewer does the merge. The merge will automatically delete the branch in gitlab, but locally the coder has to remove the branch manually. Cheers, Eric

added 2 commits

• a2082192 - Fixed errors from EK review of merge request 4
• 34845475 - Test benches for lofar2_unb2b_adc and revisions

Compare with previous version

1-4 OK, new comments: 5) Use parent library lofar2_unb2b_adc_lib instead of work in revision. Remove parent ../../src/vhdl sources from revision hdllibg.cfg. 6) Remove tb_lofar2_adc_unb2b_full.vhd?

added 2 commits

• ff37fc54 - Remove incorrectly named file. Use lofar2_unb2b_adc_lib in revisions
• ce4e177f - Add entry for jesd204b number of transceiver channels

Compare with previous version

added 37 commits

• ce4e177f...942d8f38 - 36 commits from branch master
• d129e115 - Merge branch 'master' into stat-313

Compare with previous version

merged

mentioned in commit 515222a1